The roll-off further up emphasizes this even more. The speaker sounds a lot like upper midrange only, to exaggerate quite a bit but just to give you the idea. Stories have circulated over the years that Peter Walker pushed the 2–3kHz region on purpose because he thought that sounded better and reduced the lower mids to avoid any possibility of “boom.” If so, I must say I disagree—the speakers sound better if you take this projection further away and fill in the lower mids—quite a bit better, actually, to my ears. But it’s worth noting that Walker has not been involved in design work for many decades, and obviously had no hand in the 2812.
This overall slightly peculiar balance is a serious matter for large-scale music and, once you start to notice it, for other music as well. The orchestra sounds somehow miniaturized, with diminished fullness in the brass and weakness in the cello and viola sections, and even the lower part of the violins is affected. (The bottom note of the violin is just under 200Hz.) The orchestra sounds a bit as if the instruments were toys. Perhaps this seems like the famous cartoon from the old days with the frantic audiophile and the caption, “Crisis: the second oboe is too weak.” But this is actually a real problem for orchestral or other large-scale music, and indeed for music in general. You might like it but it is wrong. In spite of the impressive bass, impressive for a relatively small full-range electrostatic anyway, the effect is miniaturization and a balance that is just not correct.
I am well aware that this type of balance is admired, perhaps even preferred, by some people—witness the ongoing popularity of the LS3/5a, which is like this to an extreme (and does not even have bass, which the Quads do have). But still, it is a problem if you want to hear music as it is (or even as it is recorded).
There is also a roll-off in room sound in the top octave or so, though this is arguably not all that unnatural in musical terms since concert hall sound is similarly rolled-off. But it sets off the 2kHz emphasis even more, by contrast, and of course affects micro-detail directly. On the harp recording mentioned, you will surely notice that the micro-structure of the sound, the little twanginess and texture that the harp has in reality, is considerably reduced compared to a speaker with a flat and non-beaming top end. Better too little top than too much—too little being at least more or less consistent with live acoustic music. But one does miss some texture with the Quads.
The Quads sound smooth and in one sense uncolored—the sense of the absence of box colorations and resonances (except in the bass). But they are not really flat in-room. They are quite smooth as speakers go but not really neutral. (This is obvious on pink noise where one can hear both the lower midrange hole and 2kHz whooshing away at you, and to a lesser extent 4–5kHz.)
Also, one needs to damp the wall behind the speakers a lot to get the best results. Dipole operation in the lower frequencies is a good way to interact with a room, but in the higher frequencies it definitely is not, from my viewpoint anyway. One really does not want the top end (what there is of it) to bounce hard off a back wall. Damping behind is really needed for satisfactory results. A lot of damping. If you do that right, however, you will be rewarded with a really extraordinary purity of sound. You have to do some work but the rewards are great.
The Quads also have a certain subtle glaze to the sound when there is a hard attack, though the present version does this much less than the original ELS63. This effect is subtle, especially if the speakers are eq’d to be balanced correctly. But on piano attacks there is a hint of glare. The speakers are never really convincing on piano music in the upper registers for this reason. The phase linearity ought to help to make piano attacks right, but somehow this does not happen. Where this comes from is hard to say. Low-level chaotic motion of the diaphragm has been suggested (the “waterfall” looks pretty junky in the lower treble), though real evidence is missing. Listen for this in any case. Many people do not seem to notice. Some people find it annoying.
In any case, if you are a fan of large music in any sense, something needs to be done about the much less subtle attenuation of the lower mids. Adding subwoofers is a good idea for bass extension and dynamics, but it will not address the lower midrange “energy hole,” which is too far up to be subwoofered away.
Enter DSP correction. The Quads are real naturals for this: The speakers can be made well balanced for large-scale music in this way, while retaining their purity, coherence, and low distortion. The result is an exceptional speaker. Some might want to add a super-tweeter for wider dispersion of the top end, and, as noted, subwoofers for the deep bass are good. But the Quads corrected to fill in the energy hole and reduce the mids above a bit are something special without supplementation at either end. This is not to say that they do not have wonderful virtues without the correction, but with the balance straightened out they enter a rarified realm of excellence. (I should add that I believe virtually any loudspeaker can be improved with DSP correction.)
People, serious people, have noticed this potential. Tam Henderson of Reference Recordings, who has access to pretty much anything, is using subwoofered Quads corrected with Lyndorf Audio’s RoomPerfect with great satisfaction. He calls it his “last system.” And engineer Tony Knight, who has taken on the problem of literal acoustical replication of signal in a room in a profoundly dedicated way, is using subwoofered and DSP-corrected Quads 989s. He is currently using Uli Bruggemann’s Acourate correction system. His recordings of recordings demonstrations at RMAF have created something of a sensation—and no wonder. Even after several generations of recordings of recordings of recordings, the resemblance to the original is striking. (Each channel is recorded with a microphone at the listening position and then the individual channels combined into a stereo recording. This sounds surprisingly like the original recording played, showing how well the system is reproducing acoustically its electrical input.)
I mention this not to emphasize the need of the Quads for DSP correction—after all, many speakers need it at least as badly—but to emphasize how well it works for Quads. Billie Jean King once described a high-bounced tennis ball as “aching to be slugged” with a powerful overhead. In a similar fashion, the Quads, with their all but incomparably low distortion and coherence, are aching to be corrected because the promise of perfection is so strong. And a DSP-corrected Quad setup comes about as close to perfect on its own terms—that is, its radiation pattern choice—as one is likely to get.